Progress and challenges in model risk management

Model risk is a relatively new focus area for risk management. In 2016, we conducted a small study into the topic, which showed that firms are at different stages of evolving practice. Following the renewed interest in the Analytics Working Group, we conducted a follow-up study to see how and in which areas the industry has advanced.

The report looks at the progress that has been made in the industry and identifies new areas of convergence, as well as the challenges that firms face today. The focus areas include key aspects of model risk management for models that require regulatory approval, including: data and definitions; quantification and capital; frameworks and governance; risk appetite and decision-making; and common challenges.

Key findings

Model risk gaining attention as a standalone risk type

The importance of model risk and model risk management is confirmed by the fact that firms are increasingly considering the risk as a principal risk type, and not just a sub-type of operational risk.

Firms face the challenge of establishing a framework while the scope keeps expanding

Eighty-six per cent of respondents stated that they have established a group-wide or over-arching framework for model risk management. Nevertheless, extending or improving the framework is high on the agenda for many – and constitutes one of the biggest challenges.

Figure 1. Areas of progress and development in model risk management

Model risk management shows areas of progress and convergence

Many of the areas where we expected to see industry convergence show considerable progress. The responsibilities for model approval are now more clearly defined. Also, the primary roles and responsibilities in the model lifecycle show broad alignment across jurisdictions.

There are opportunities in using model risk information for enterprise risk management and decision-making

As frameworks become more established, one of the next milestones for many firms will be establishing and tracking model risk appetite and using the information in decision-making. Seventy-seven firms are already including model risk information in senior management reporting, but only 58 per cent have a risk appetite statement in place.

Quantifying model risk depends on elements that firms are still working towards

As only 48 per cent of study participants are currently required to quantify model risk capital, it does not come as a surprise that many firms are missing important elements for model risk quantification. As firms come to grips with the assessment of model risk and its quantification, the potential benefits from being able to use this information for internal reporting and decision-making will increase.